Analytica Chimica Acta 817(2014)23–27Contents lists available at ScienceDirectAnalytica ChimicaActaj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /a caA new interface for coupling solid phase microextraction with liquid chromatographyYong Chen ∗,Leonard M.SidiskySupelco,595North Harrison Road,Bellefonte,PA 16823,USAh i g h l i g h t s•A new solid phase microextraction(SPME)–liquid chromatography (LC)interface.•Fiber desorption occurred off-line,but all desorption solvent could be conveniently injected into LC systems with the interface.•The new SPME–LC interface was robust and reliable for coupling SPME with LC for both qualitative and quan-titative analysis.g r a p h i c a la b s t r a cta r t i c l ei n f oArticle history:Received 3December 2013Received in revised form 15January 2014Accepted 26January 2014Available online 6February 2014Keywords:Solid phase microextraction (SPME)InterfaceLiquid chromatography (LC)Polycyclic aromatic hydrocarbons (PAHs)a b s t r a c tA modified Rheodyne 7520microsample injector was used as a new solid phase microextraction (SPME)–liquid chromatography (LC)interface.The modification was focused on the construction of a new sample rotor,which was built by gluing two sample rotors together.The new sample rotor was further reinforced with 3pieces of stainless steel tubing.The enlarged central flow passage in the new sample rotor was used as a desorption chamber.SPME fiber desorption occurred in static mode.But all desorption solvent in the desorption chamber was injected into LC system with the interface.The ana-lytical performance of the interface was evaluated by SPME–LC analysis of PAHs in water.At least 90%polycyclic aromatic hydrocarbons (PAHs)were desorbed from a polyacrylonitrile (PAN)/C18bonded fuse silica fiber in 30s.And injection was completed in 20s.About 10–20%total carryovers were found on the fiber and in the interface.The carryover in the interface was eliminated by flushing the desorption chamber with acetonitrile at 1mL min −1for 2min.The repeatability of the method was from 2%to 8%.The limit of detection (LOD)was in the mid pg mL −1range.The linear ranges were from 0.1to 100ng mL −1.The new SPME–LC interface was reliable for coupling SPME with LC for both qualitative and quantitative analysis.©2014Elsevier B.V.All rights reserved.1.IntroductionSolid phase microextraction (SPME)is a convenient and rapid sample preparation technique [1–3].It has being extensively cou-pled with gas chromatography (GC)for the analysis of volatile organic compounds (VOCs)and semi-VOCs [4,5].This unique technique integrates sampling,sample preparation,and sample∗Corresponding author.Tel.:+18143595914;fax:+18143595702.E-mail address:yong.chen@ (Y.Chen).introduction into a single step,and enables complete automatic SPME–GC analysis [6].SPME is also coupled with liquid chromatog-raphy (LC)for the analysis of semi-VOCs,non-volatile compounds,and thermal liable compounds [7].SPME–LC is different from SPME–GC in that the desorption in SPME–LC is solvent desorption while it is thermal desorption in SPME–GC.SPME–LC desorption can be done either off-line or on-line.Off-line desorption does not require special interfaces.A SPME fiber is immersed into a desorption solvent contained in a vial to desorb the extracted analytes.The desorption solvent containing the desorbed analytes is then injected into a LC system.Off-line0003-2670/$–see front matter ©2014Elsevier B.V.All rights reserved./10.1016/j.aca.2014.01.05624Y.Chen,L.M.Sidisky/Analytica Chimica Acta817(2014)23–27desorption is versatile.But one disadvantage is that not all the des-orbed analytes are injected into LC systems.Jinno et al.developed off-line SPME–LC interface and its improvement[8,9].These inter-faces facilitated the injection of a portion of desorbed analytes into LC system.On-line SPME–LC desorption has to be done in a special interface,which is similar to GC injection ports for SPME–GC thermal desorption.However,on-line SPME–LC coupling is much more challenging than SPME–GC coupling due to the nature of solvent desorption and the high pressure presented in LC systems.Chen and Pawliszyn developed thefirst SPME–LC on-line inter-face[10].The interface included a desorption chamber and a Rheodyne7161injection valve.The desorption chamber was a stainless steel tee joint.The side outlet and the bottom outlet of the tee joint were connected to the injection valve,and the upper outlet of the tee joint was used to introduce a SPMEfiber device.The seal of the SPME device was done by a piece of poly(ether ether ketone)(PEEK)tubing and a two-piecefinger-tight PEEK union.It was claimed that the seal could withstand pressure up to4500psi.When the injection valve was in“load”position,the tee joint was at ambient pressure so that SPME fibers could be introduced into or moved out of the inter-face.When the injection valve was in“injection”position,the mobile phaseflew through the tee joint and carried the des-orbed analytes to the LC column.The interface was evaluated with SPME–LC analysis of polyaromatic hydrocarbons(PAHs)in water,and the performance was validated with standard loop injec-tion.Boyd-Boland and Pawliszyn improved the above on-line SPME–LC interface[11].Thefirst modification they did was that a larger(0.75mm inner diameter(i.d.))desorption chamber was used to accommodate swollenfibers.The second modification was that a0.4mm i.d.GC ferrule and a connector replaced the PEEK tubing and the PEEK union to seal around the inner tubing of the SPMEfiber assembly.The new seal mechanism provided more reli-able seal and reduced the chance offiber damage.The interface was evaluated with the analysis of non-ionic surfactant Trion X-100in water.The performance of severalfiber coatings was compared.It was demonstrated that at least90%desorption efficiency could be achieved in1min.Rodrigues et al.developed a heated SPME–LC interface[12]. Although the interface was able to perform on-line desorption, the critical sealing mechanism was not discussed.It seemed that the improvement was only focused on the heat assisted desorption.All the above on-line interfaces suffer from leak andfiber coating damage.Lord reviewed state-of-the-art of interfacing SPME with LC [13].The challenges for on-line coupling SPME with LC were sum-marized as availability of commercialfibers,sealing mechanism, desorption optimization,and automation options.Recently,Chen and Sidisky improved the commercial SPME–LC interface[14].The commercial interface was based on Boyd-Boland and Pawliszyn’s design.Both SPMEfiber assembly and the desorp-tion chamber were modified to address the issues of leak andfiber coating damage.Thefirst modification was the use of polyacryloni-trile(PAN)/C18coating which does not swell in organic solvents. The second modification was the use of a solid innerfiber support with a much larger outer diameter(o.d.)than the o.d.of thefiber coating.The third modification was the enlargement of the hole of the ferrule to accommodate the innerfiber support.Another seal-ing mechanism by the use of a special designed PEEKfitting was also presented in the research.It was demonstrated that the prob-lems of leak andfiber coating damage were effectively eliminated with the improved interfaces.In this research,a new SPME–LC interface was developed and tested.The goal was to improve the applicability of SPME–LC anal-ysis.2.Experimental2.1.Chemicals and materialsPolycyclic aromatic hydrocarbon(PAH)525mixture(500␮g mL−1of acenaphthylene,fluorene,phenanthrene,anthracene, pyrene,benz[a]anthracene,chrysene,benzo[b]fluoranthene, benzo[k]fluoranthene,benzo[a]pyrene,dibenzo[ah]anthracene, indeno[l,2,3-cd]pyrene,and benzo[ghi]perylene prepared in methylene chloride),LC water,Acetonitrile(ACN),Epoxy glue, were obtained from Aldrich(Milwaukee,WI,USA).Rheodyne 7520Microsample injector,sample rotor for Rheodyne7520 Microsample injector,SPMEfibers,Fiber holders for automatic sampling,micro syringes,and2mL of vials were obtained from Supelco(Bellefonte,PA,USA).Gauge23TW and19TW Stainless steel tubings were purchased from Vita needle(Needham,MA, USA)2.2.SPME–LC interface deviceThe SPME–LC interface was a modified Rheodyne7520 Microsample injector(Fig.1).The major modification was done on the rotor.The largest i.d.of the sample chamber in the original rotor was about0.3mm,and the height of the rotor was7mm.The size of the sample chamber was considerably small to accommodate a standard commercial SPMEfiber.A new rotor was built by gluing two rotors together.The two holes on each outer side were enlarged to i.d.0.635mm.A piece of SS tubing(o.d.0.635mm,i.d.0.432mm, height12mm)was inserted into each of the holes.The two pieces of tubing were permanentlyfixed in the holes with epoxy glue.The central hole in the rotor was enlarged to i.d.1.07mm.A piece of SS tubing(o.d.1.07mm,i.d.0.812mm,height12mm)was inserted into the hole.The tubing was permanentlyfixed in the holes with epoxy glue.The new rotor replaced the original rotor.Fig.1shows the cut-away schematic view of valveflow path for the SPME–LC interface.2.3.SPME proceduresAutomatic SPME extraction was performed on a CTC Combi PAL autosampler(CTC Analytical,Zwingen,Switzerland).The SPME fibers were pre-conditioned in ACN for15min and air-dried for 30s prior to thefirst extraction.The standard solutions were prepared by diluting PAH525mixture with methanol.The con-centrations of the standard solutions ranged from0.1␮g mL−1 to100␮g mL−1.The working solutions were prepared by spiking the standard solutions into1.4mL water contained in2mL vials. Immersion SPME extraction was performed with the following parameters:incubation/extraction temperature:50◦C,incubation time:5min,extraction time:20min,agitation rate:500rpm.SPME fiber:PAN/C18fiber.The length of thefiber was7mm.After extraction,the SPMEfiber needle was introduced into the sample needle port in the interface.Then thefiber was exposed and immersed into the desorption solvent(ACN)(Fig.1A).After 30s desorption at room temperature,thefiber was withdrawn into the outer needle of thefiber assembly,and thefiber assembly was removed from the sample needle port.Immediately after the removal of thefiber assembly the knob of the interface was rotated 45◦to switch the interface from“LOAD”mode to“INJECT”mode (Fig.1B).The mobile phase from LC pumpflew through the desorp-tion chamber and carried the desorbed analytes in the desorption chamber to the LC column for separation.A Series III LC pump(LabAlliance,state college,PA,USA)was used to deliver solvent(ACN)into the desorption chamber for cleanup and desorption.Y.Chen,L.M.Sidisky/Analytica Chimica Acta817(2014)23–2725Fig.1.The SPME–LC interface in the(A)load and(B)inject mode.2.4.InstrumentationAgilent1200LC system with data acquisition Chemstation for LC3D systems(Rev.B03.02)was utilized for the analysis.The separation was carried out on an Ascentis Express C18column (2.7␮m,150mm×4.6mm)obtained from Supelco.The mixture of acetonitrile(ACN)and water(15/85,v/v)was used as mobile phase A.ACN was used as the mobile phase B.The chromatographic conditions were:the column was maintained at25◦C.From0to 0.5min,isocratic75%B,0.5mL min−1;0.5–4min,gradient75–90% B,0.5–1mL min−1;4–10min,gradient90–100%B,1mL min−1; 10–15min isocratic100%B,1mL min−1;15–15.01min,gradient 100–75%B,1mL min−1;15.01–19min,isocratic75%B,1mL min−1; 19–19.01min,isocratic75%B,1–0.5mL min−1;19.01–20min iso-cratic75%B,0.5mL min−1.The effluent was monitored by UV(Ultra Violet)detection at254nm.3.Results and discussion3.1.The interfaceThe volume of the desorption chamber in the rotor was esti-mated at about8␮L.it is about32times larger than the volume of thefiber coating and17times larger than the displacement volume of thefiber.The dimensions of the desorption chamber ensured that first,thefiber could freely move in and out the chamber without being damaged.Second,the volume of the desorption solvent was large enough to ensure efficientfiber desorption.In the meantime, it was not large enough to cause significant peak distortion.When thefiber was exposed into desorption solvent,the desorp-tion occurred in static mode.After desorption,thefiber was withdrawn into the outer needle prior to injection.Since thefiber was not exposed to the mobile phase at any time,the interface should be regarded as an off-line SPME–LC pare to other off-line SPME–LC interfaces[8,9],it is possible to inject all desorbed analytes into the LC column with this interface,in a similar fashion to on-line SPME–LC pare to on-line SPME–LC interfaces[10–12],SPMEfiber assemblies did not have to withstand high pressure with this interface,so that coating damage and the leak of mobile phase fromfiber assembly and from the seal around thefiber assembly at high pressure associated with on-line interfaces were avoided.Whenever the interface was in“load”or “inject”mode,it was the interface that was subjected to high pres-sure.Since the new rotor was built with two original rotors and reinforced with SS tubing,no leak was observed when the inter-face was subjected to the pressure test performed at350bar for 30min.The same interface had been used for more than300times and it was still leak-free.3.2.Evaluation of the interfacePAHs were selected as the test compounds to evaluate the ana-lytical performance of the interface.PAN/C18was thefiber coating used to extract PAHs from water samples.The coating had high affinity to PAHs,and it did not swell in most organic solvents at room temperatures[14].The SPME extraction parameters were the same as those reported in the reference[14],the only exception was that the extraction time was shorten to20min,which was the same as the LC run time.Since the volume of desorption solvent was small(about8␮L),fast chromatographic separation with satisfied peak shape and resolution was possible.LC parameters were opti-mized,and the separation of the PAHs wasfinished in10min.Fig.2 shows a typical chromatogram for the SPME–LC analysis of PAHs in water.Faster separation was possible but not explored,because the rate-limiting step was the SPME extraction.The interface was investigated for its specific function of desorp-tion and injection.The desorption of PAHs from PAN/C18fiber was discussed and investigated in the literature[14].For the specific interface,desorption time profiles were determined to investigate the pattern that the desorption efficiency changed with desorption time.Fig.3demonstrates that there was no increase in the amount of PAHs desorbed from the PAN/C18fiber after30s desorption.The results agreed with previous study that the desorption of PAHs from26Y.Chen,L.M.Sidisky /Analytica Chimica Acta 817(2014)23–27Fig. 2.A chromatogram obtained with SPME–LC analysis of 5ng mL −1PAHs in water.Peak identification:(1)fluorene,(2)phenanthrene,(3)anthracene,(4)pyrene,(5)benz[a ]anthracene and chrysene (6)benzo[b ]fluoranthene,(7)benzo[k ]fluoranthene,(8)benzo[a ]pyrene,(9)indeno[l,2,3-cd ]pyrene.PAN/C18fiber in ACN was a very fast process [14].After desorption,all the solution in the desorption chamber was injected into the LC system.The injection time profiles were determined and presented in Fig.3b,which demonstrates that there was no increase in the responses of PAHs after 20s injection.Thus,30s desorption time and 20s injection time were used for subsequent analysis.Though optimization of desorption and injection parameters resulted in maximum responses of PAHs,it did not guarantee com-plete recovery of extracted PAHs in the fiber coating.Investigation of the carryovers and mass conservation was complementary to the optimization of desorption and injection processes.For static(a)(b)20406080100120140300250200150100500P e a k A r e a (m A u )DesorpƟon Time (sec)20406080100120140300250200150100500P e a k A r e a (m A u )Injec Ɵon Time (sec )Fig.3.Optimization of desorption efficiency.(a)Desorption time profiles,(b)Injec-tion time profiles. :fluorene, :phenanthrene, :anthracene.Table 1Carryovers of PAHs.Carryover (%)In the interfaceOn the fiberFluorene8.1 4.9Phenanthrene 8.4 5.0Anthracene 8.3 5.0Pyrene9.4 6.3Benzo[b ]fluoranthene 11.27.9Benzo[k ]fluoranthene 11.68.1Benzo[a ]pyrene8.57.9Indeno[1,2,3-cd ]pyrene8.19.4Note :The UV responses of acenaphthylene,dibenzo[ah ]anthracene,and benzo[ghi ]perylene were too weak to be reliably quantified.Benz[a ]anthracene and chrysene could not be separated completely.So their results were not reported throughout the research.desorption,carryover in the fiber coating was almost unavoidabledue to partition of analytes between the desorption solvent and the fiber coating.This problem could be alleviated by the use of large volume of desorption solvent with strong affinity toward analytes.In this case,the desorption solvent was ACN,and its volume was 32times larger than the volume of the fiber coating.The carry-overs of PAHs in the fiber coating were determined by desorbing the fiber into 100␮L ACN for 30s immediately after the first desorp-tion and injecting the ACN into the LC for quantification.It was found that the carryovers of PAHs ranged from 5to 9%(Table 1).The carryovers found in this study were generally larger than the carryovers of PAHs in the fiber coating reported by the use of com-mercial desorption chamber.The reason was that the volume of ACN in the commercial desorption chamber was 75␮L,which was significantly larger than the volume of ACN in the modified Rheo-dyne 7520desorption chamber (8␮L).Increase of the volume of the desorption chamber would increase the volume of the desorp-tion solvent,which would reduce the carryovers of PAHs on the fiber.But the disadvantage of the use of large volume of desorp-tion solvent was that optimization of LC separation was much more challenging.To ensure that the fiber was clean for the next extrac-tion,the fiber was immersed into 100␮L ACN for 30s after the fiber had been desorbed in the interface.No carryover on the fiber was detected after the cleanup.Another source of carryover was the carryover in the interface,which was caused largely by incomplete transfer of desorbed ana-lytes onto LC column.Theoretically it only took about 1s to replace all the 8␮L desorption solvent with the mobile phase at a flow rate of 0.5mL min −1.Practically,it took much longer than 1s to replace all the desorption solvent due to the parabolic flow of the mobile phase in the desportion chamber and insufficient agitation of the desorption solvent located in the edge of the base of the desorption chamber.The carryovers of PAHs in the interface were determined with the following procedure.When the first desorp-tion and injection had been done,the cleanup of the interface was not performed.After the first separation had been done,the inter-face was subjected another injection with the continuous flow of the mobile phase through the desorption chamber to determine the carryovers of PAHs in the interface.It was found that the car-ryovers of PAHs ranged from 8to 12%.The carryovers of PAHs in the interface could be reduced with prolonged injection time.But it would cause tailing peaks and raised baseline.In addition,it was learned from the injection time profiles that the responses of PAHs did not change significantly with injection time.After optimization of desorption and investigation of the car-ryovers,mass conservation of PAHs during SPME extraction and desorption was checked.Phenanthrene was used as the exam-ple.Its total mass in the sample was 7ng.Its mass in the sample after SPME extraction was determined as 3.3ng.So the mass ofY.Chen,L.M.Sidisky/Analytica Chimica Acta817(2014)23–2727 Table2Summary of calibration results for the SPME–LC analysis of PAHs in water.LOD(ng mL−1)Repeatability(%,n=7)Linearitya b R2Linear range(ng mL−1)Fluorene0.020 1.8 2.73750.7130.99980.2–100Phenanthrene0.049 2.19.7905 2.04020.99990.2–100Anthracene0.033 1.820.71218.6780.99780.1–100Pyrene0.28 3.1 1.13920.021610.5–100Benzo[b]fluoranthene0.11 4.2 5.22220.79990.99680.2–20Benzo[k]fluoranthene0.18 4.9 4.3947−0.44190.98960.2–5Benzo[a]pyrene0.158.4 2.64720.11550.98970.2–10Indeno[1,2,3-cd]pyrene0.16 6.1 3.1834−0.58780.99440.2–5phenanthrene removed from the sample was3.7ng.The recovered mass of phenanthrene from the SPMEfiber was the sum of the mass from thefirstfiber desorption(3.1ng),the mass from the interface carryover(0.3ng),and the mass from thefiber carryover(0.18ng). So the recovered mass of phenanthrene was3.58ng,which was about96.8%of the mass of phenanthrene removed from the sample. This demonstrated that the analytical processes associated with extraction,desorption,injection,and separation were reliable.The analytical performance of SPME–LC analysis of PAHs by the use of the interface was evaluated.The repeatability was deter-mined by analyzing7PAHs aqueous solutions at a concentration of 5ng mL−1with the SPME–LC method.The relative standard devia-tion of PAHs peak areas ranged from2to9%.The limit of detection (LOD)was estimated as the concentration from which the result-ing chromatographic peak had signal-to-noise(S/N)ratio of3.The linearity of the method was estimated by SPME extraction of work-ing PAHs solutions with concentrations ranging from0.1ng mL−1 to100ng mL−1.The standard solutions were prepared in triplicate for each concentration.Table2summarizes LOD,repeatability,and linear range for each analyte.From the results it can be concluded that the SPME–LC interface was suitable for both qualitative and quantitative analysis of PAHs in water.Compare to the SPME–LC interface improved in the previous study[14],the SPME–LC interface developed in this research had some advantages and disadvantages.First,the solvent desorption of SPMEfiber in the SPME–LC interface was performed only in static mode,while it could be done in static and dynamic modes in the previous interface.Second,the volume of the desorption chamber in the SPME–LC interface was smaller than that in the previous interface.So it was easier to optimize the desorption and separa-tion parameters,and separation could be done in shorter time.In addition,high strength solvents could be used for desorption.The disadvantage was that the carryover in thefiber was more promi-nent with the small volume of the desorption solvent.Third,the SPME–LC interface was more robust than the previous one because the only consumable part in the interface was the sample rotor. The same sample rotor had been used for more than300times,and no leak was observed.In the meantime,the interface was easier to operate because the only operation with the interface was the rotation of the knob.Forth,there were more carryovers with the SPME–LC interface than those found in the previous interface.The more carryovers in the interface might be ascribed to that there was adsorption of PAHs on the inner surface of the desorption chamber of the SPME–LC interface.It was observed that the adsorption of PAHs on the inner surface of the desorption chamber was signifi-cantly reduced after the SS tubing was inserted in the desorption chamber.However,the whole desorption chamber surface could not be replace with SS material due to the requirement for seal. One improvement for the SPME–LC interface is that the rotor be made of PEEK(polyether ether ketone),so that the interface car-ryover would be reduced.The more carryovers in the SPMEfiber might be ascribed to that thefiber was only subjected to static desorption,while a brief dynamic desorption after static desorp-tion were used in the previous interface.Fifth,analytical validation by SPME–LC analysis of PAHs in water demonstrated that these two interfaces had comparable analytical performance in terms of desorption efficiency,carryover,recovery,repeatability,LOD,and linearity.4.ConclusionThe newly developed SPME–LC interface was capable of cou-pling SPME with LC by off-line desorption and on-line injection. 90%or more desorption efficiency was achieved for the desorption of PAHs from the PAN/C18fiber,and the desorption process was as short as30s.On-line injection allowed injection of all desorp-tion solvent,and the injection process was as short as20s.The total carryovers of PAHs on thefiber and in the interface were from 10%to20%.The carryover in the interface was eliminated by clean-ing the desorption chamber with ACN at1mL min−1for2min.The carryover on thefiber was eliminated by desorbing thefiber in 100␮L ACN for30s.Rapid LC separation of PAHs was achieved in 10min owing to the small volume of the desorption solvent.The SPME–LC analytical procedure for the analysis of PAHs in water with the SPME–LC interface was reproducible,and had low LOD and wide linear ranges.It was,therefore,demonstrated that the SPME–LC interface was suitable for coupling SPME with LC for both qualitative and quantitative analysis.References[1]C.L.Arthur,J.Pawliszyn,Anal.Chem.62(1990)2145.[2]H.Lord,J.Pawliszyn,J.Chromatogr.A885(2000)153.[3]J.Pawliszyn,Anal.Chem.75(2003)2543.[4]J.Pawliszyn,Solid Phase Microextraction–Theory and Practice,Wiley-VCH,New York,1997.[5]J.Pawliszyn(Ed.),Applications of Solid Phase Microextraction,RSC,Cornwall,UK,1999.[6]J.O’Reilly,Q.Wang,L.Setkova,J.P.Hutchinson,Y.Chen,H.L.Lord,C.M.Linton,J.Pawliszyn,J.Sep.Sci.28(2005)2010.[7]J.Pawliszyn(Ed.),Sampling and Sample Preparation for Field and Laboratory,Elsevier,Amsterdam,2002.[8]K.Jinno,T.Muramatsu,Y.Saito,Y.Kiso,S.Magdic,J.Pawliszyn,J.Chromatogr.A754(1996)137.[9]K.Jinno,M.Taniguchi,M.Hayashida,J.Pharm.Biomed.Anal17(1998)1081.[10]J.Chen,J.Pawliszyn,Anal.Chem67(1995)2530.[11]A.A.Boyd-Boland,J.B.Pawliszyn,Anal.Chem.68(1996)1521.[12]J.C.Rodrigues,o,C.Fernandes,C.Alves,A.S.Contadori,ncas,J.Chromatogr.A1105(2006)208.[13]H.L.Lord,J.Chromatogr.A1152(2007)2.[14]Y.Chen,L.Sidisky,Anal.Chim.Acta743(2012)61.。